VOGONS


Need help with my first 486 boards.

Topic actions

Reply 20 of 35, by Kubik

User metadata
Rank Member
Rank
Member

No video would be one long, followed with two shorts. I'd recommend to use a diagnostic card, those are so cheap on Aliexpress that it's pretty much a no-brainer, and I wouldn't bother touching the board again before having a diag card at hand 😀
You probably want the one with both ISA and PCI: https://www.aliexpress.com/item/Free-Shipping … 2625200469.html

Reply 21 of 35, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
BeginnerGuy wrote:
Sometimes that happens, that's why I said I'm guessing whether the video card plugged in or not it still wont beep --> it's prob […]
Show full quote
Ozzuneoj wrote:

I get no beeps when RAM is inserted. The only time I get beeps is when the CPU is installed without RAM.

I have tons of boards to play with so I'm planning to try a hot swap program for a BIOS chip, just to test. I read about it last night and it sounds fairly simple. I checked all of my really old network cards and modems that use 32pin ROM chips and it doesn't look like any have the A17 pin connected which is needed for flashing. I'll experiment with it today if I can. Fighting a nasty head cold right now. -_-

Sometimes that happens, that's why I said I'm guessing whether the video card plugged in or not it still wont beep --> it's probably hanging up at RAM during post and not progressing on to video. Usually bad or no video on these boards will result in beeps. That's just my experience from building my own socket 3 setup.. your mileage may vary.

Let us know if you have any luck. I have a few boards with fried bios ROMs and I've been far too cheap to invest in an EEPROM burner to fix em, would be awesome if I could do it cheaply 😒

Head colds are going around.. I was in bed for almost 2 weeks in November with one.. Now I'm avoiding all humans until spring 🤣

Since I'm not 100% sure that its the BIOS, it looks like I'll just have to wait until I get some FP memory to try in this board. I don't want to spend all day messing with it if its a RAM problem. I have one more machine I can tear into to see if it has FP RAM but I'm 99% sure it has EDO since it is one of my DOS\9x systems with an Intel 430HX chipset and P 200MMX.

Its pretty hard to find dirt cheap (I mean DIRT CHEAP) FP RAM since it doesn't seem to look any different and many people just call any 72pin RAM "EDO". I don't really want to spend $8 on RAM to test a board that could be dead, considering it may be the only time I'll need 72pin FP SIMMs.

Also, is there likely to be any difference between what I'd need for a board like this and RAM advertised as being for\from an "Apple"? Does it need to be Parity RAM?

Now for some blitting from the back buffer.

Reply 22 of 35, by Kubik

User metadata
Rank Member
Rank
Member

FPM RAM can be easily identified by the chips. As long as you can recognize the chip marking on that eBay picture, you're fine 😀
BTW: so far, I didn't really care about EDO/FPM. I didn't have the chance to actually get my hands on board that wouldn't accept EDO, so I'm putting EDO memory everywhere (including Atari ST and Macs). The chance it's EDO/FPM problem in your case is rather low. Did you try to remove/disable cache?
EDIT: and no, it doesn't need to be parity SIMM.

Reply 23 of 35, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
Kubik wrote:

FPM RAM can be easily identified by the chips. As long as you can recognize the chip marking on that eBay picture, you're fine 😀
BTW: so far, I didn't really care about EDO/FPM. I didn't have the chance to actually get my hands on board that wouldn't accept EDO, so I'm putting EDO memory everywhere (including Atari ST and Macs). The chance it's EDO/FPM problem in your case is rather low. Did you try to remove/disable cache?
EDIT: and no, it doesn't need to be parity SIMM.

Thanks for the tips. I'll try removing cache next. I did order an ISA\PCI diagnostic card. We'll see if its helpful or not. 🤣

Now for some blitting from the back buffer.

Reply 24 of 35, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

Well, I tried pull two of the cache chips, then pulling the other two and swapping them out. No difference. I don't have another chip to replace the TAG, but I pulled it and tried running without it, also with no change. The board only has settings for 128,256 and 512k cache, nothing about disabling it.

I closely inspected the bottom of the board and found a broken solder joint on the CPU socket. I thought I'd just figured out the secret of the universe, but after reflowing the joint it still didn't make a difference.

I'm going to shelve this board until I get my diagnostic card... there are just too many variables at this point, because it feels like I've covered all of the basics.

Now for some blitting from the back buffer.

Reply 26 of 35, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

Okay, I got my POST test card.

Here are the results:
(+ indicates the part is installed, - indicates removed)

+CPU +RAM +VGA = 0503 or 0504 (seems to be random as to which it displays each time)
+CPU +VGA -RAM = C7C6 (long beeps)
+CPU +RAM -VGA = 0503 or 0504
+CPU +RAM, -VGA -All Cache = 0503 or 0504
+CPU +RAM, -VGA -All Cache -AMIKEY-2 Controller = 0503 or 0504

I bought this card:
https://www.amazon.com/Optimal-Shop-Computer- … ds=pc+post+card

Can't find the link for the manual online, but its available.

Anyone have any ideas?

Now for some blitting from the back buffer.

Reply 27 of 35, by Kubik

User metadata
Rank Member
Rank
Member

Well, I am not sure if I have the right AMI BIOS code table, but 03 is BIOS checksum, 04 is keyboard controller. That would be a good starting point. However, the fact it's randomly alternating is worrying me, however, the card seems to have some sort of intelligency on its own.
Typically, you ignore the upper half of the display, and only read the lower half (long story...), however, this looks like it does more, perhaps remembering the previous POST code. If that's the case, we might be actually getting 05 which probably means chipset init - and in that case, I'd pronounce the board dead...
Anyway - replacing the keyboard controller and BIOS chip is something you should try. I can imagine some bits in the EPROM flipped over the years...
Do you have manual for the board?

Reply 28 of 35, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
Kubik wrote:
Well, I am not sure if I have the right AMI BIOS code table, but 03 is BIOS checksum, 04 is keyboard controller. That would be a […]
Show full quote

Well, I am not sure if I have the right AMI BIOS code table, but 03 is BIOS checksum, 04 is keyboard controller. That would be a good starting point. However, the fact it's randomly alternating is worrying me, however, the card seems to have some sort of intelligency on its own.
Typically, you ignore the upper half of the display, and only read the lower half (long story...), however, this looks like it does more, perhaps remembering the previous POST code. If that's the case, we might be actually getting 05 which probably means chipset init - and in that case, I'd pronounce the board dead...
Anyway - replacing the keyboard controller and BIOS chip is something you should try. I can imagine some bits in the EPROM flipped over the years...
Do you have manual for the board?

Whoops, mine is an Award BIOS. Sorry!

I don't have a manual, no. Manuals for 486 boards seem few and far between online.

Now for some blitting from the back buffer.

Reply 29 of 35, by Kubik

User metadata
Rank Member
Rank
Member

Luckily, the post codes are similar, so it still points to keyboard controller 😀 Were you able to identify the board model and manufacturer?
Edit: could it be this one? http://stason.org/TULARC/pc/motherboards/T/TM … AF-VER-3-0.html

Reply 30 of 35, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
Kubik wrote:

Luckily, the post codes are similar, so it still points to keyboard controller 😀 Were you able to identify the board model and manufacturer?
Edit: could it be this one? http://stason.org/TULARC/pc/motherboards/T/TM … AF-VER-3-0.html

Yes that's the board. 😀

I posted that link earlier. I'm so glad there's at least some documentation out there.

I tried pulling the keyboard controller completely and it didn't make any difference. Is it likely to be the controller still? Does a board not POST at all without a functional keyboard controller? I'm a bit confused as to what the codes are telling me though. Does the code indicate the last step that was completed successfully or is it the one that it is hanging up on?

I will say, everything I've read online points to EDO very rarely being compatible with 486 motherboards. People talk of having to use very specific configurations of FPM SIMMs to get many 486 boards going. I did find a someone on eBay selling 16MB 4x36 Parity 60ns FPM SIMMs in lots of 10 for only $20. I made an offer and he counter-offered at $19. I had $7 in ebay bucks though, so I got 10 sticks for like $12. This configuration (a single 4x36 stick) is mentioned on that TH99 page for this board as a valid configuration, so it looks like the board requires parity RAM and most likely requires FPM based on comments I've found online. This particular ALi chipset is fairly uncommon so I don't have 100% concrete evidence for this, but I'm willing to experiment for $1.20 per stick. I can always use them to populate other boards as I inevitably come across more old 486 boards in the future. 16MB seems like a good number for a 486 anyway. I can drop some more in if necessary.

I'll try the memory when I get it. If that doesn't work, I'll keep my eyes open for a super cheap keyboard controller. Are they generally universal?

Now for some blitting from the back buffer.

Reply 31 of 35, by Kubik

User metadata
Rank Member
Rank
Member

BIOS keeps a list of tasks together with the corresponding POST code. Before certain task is started, the code is sent to the diag card, so that if the BIOS hangs in the task, the number shows what task is it hanging in.
I didn't check the sources, but I wouldn't be surprised if missing / bad keyboard controller blocked the POST from continuing. There's e.g. the so-called A20 gate implemented in keyboard controller and that's important for accessing memory.
If I am in the right mood next week, I can check what exactly is happening when keyboard controller is missing, and will eventually let you know 😀

Reply 32 of 35, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

Okay, I got some 60ns Fast Page Parity RAM and the board still hangs up with the same post code (0503 mostly). The memory has 12 chips, even though the picture showed 9, but I got 10 brand new 16MB sticks for $19, which I thought was pretty good! Being 12 chips shouldn't matter, right? As far as I can tell it is parity RAM.

The stick has eight of these: http://www.datasheetlib.com/datasheet/1344218 … iconductor.html

And two of these mounted horizontally toward the bottom: http://webcache.googleusercontent.com/search? … n&ct=clnk&gl=us

I also came into a possession of a Cyrix DX4-100 3.45v CPU as well so I tried that and it didn't make any difference either. That CPU could be bad though, as it had some discoloration on some of the pins and I got a whiff of that infamous "hot silicon" smell when I cleaned the pins off with alcohol.

Also, this is the exact manual I have:
http://dimm.me/wp-content/uploads/PC-Analyzer-manual.pdf

So, now I'm basically left either trying to reflash the BIOS or finding a replacement keyboard controller.

Anyone have any other ideas or know where I could find a compatible keyboard controller for cheap?

Now for some blitting from the back buffer.

Reply 33 of 35, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

Okay, I've made some progress! I recently received a bunch of 486 boards. Sadly almost all of them were severely damaged by battery corrosion from long term storage. I managed to salvage some components from the boards though. One was an "American Megatrends" keyboard controller... a bit older than the Amikey-2 that came in the 486 board in this thread, but I figured the worst thing that could happen is that the whole thing would go up in smoke and I could finally moved on to something else.

To my amazement, it worked! Sort of... The keyboard's LEDs light up briefly and the system passes POST and displays the BIOS screen but gives a keyboard error and no input works. I assume it's because this keyboard controller came out of a much older ISA only board.

Are keyboard controllers generally pin compatible? I have a few more from different brands I can try but I don't want to do any physical damage. I don't have another Amikey-2.

Now for some blitting from the back buffer.

Reply 34 of 35, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

Woo! Found another controller, a standard "Amikey", in an incredibly corroded 486 board. The board was so corroded in one area that the "clamps" (for lack of a better word) inside the keyboard controller socket actually came out easily with the controller chip. I brushed them off into the trash, used deoxit d5, a toothbrush and a small flathead to clean the legs on the controller, inserted it into the PCI48AF board and it works perfectly! Now the board just needs a battery of some kind. There is a circle around the battery terminals that seems to indicate that a CR2032 would work here, but I removed a standard rechargeable NiMh battery from this spot... how would I prevent a CR2032 from being charged and damaged?

Now for some blitting from the back buffer.

Reply 35 of 35, by bjwil1991

User metadata
Rank l33t
Rank
l33t

Use a blocking diode to prevent the charging circuitry from attempting to charge said battery. I looked at a 386 board that had the NiCd battery removed and it has solder spots for a 2-pin jumper to prevent the charging circuitry from attempting to charge a lithium battery or two.

Discord: https://discord.gg/U5dJw7x
Systems from the Compaq Portable 1 to Ryzen 9 5950X
Twitch: https://twitch.tv/retropcuser